Paper 2023/192

Faithful Simulation of Randomized BFT Protocols on Block DAGs

Hagit Attiya, Technion – Israel Institute of Technology
Constantin Enea, Computer Science Laboratory of the École Polytechnique
Shafik Nassar, Technion – Israel Institute of Technology
Abstract

Byzantine Fault-Tolerant (BFT) protocols that are based on Directed Acyclic Graphs (DAGs) are attractive due to their many advantages in asynchronous blockchain systems. These DAG-based protocols can be viewed as a simulation of some BFT protocol on a DAG. Many DAG-based BFT protocols rely on randomization, since they are used for agreement and ordering of transactions, which cannot be achieved deterministically in asynchronous systems. Randomization is achieved either through local sources of randomness, or by employing shared objects that provide a common source of randomness, e.g., common coins. A DAG simulation of a randomized protocol should be faithful, in the sense that it precisely preserves the properties of the original BFT protocol, and in particular, their probability distributions. We argue that faithfulness is ensured by a forward simulation. We show how to faithfully simulate any BFT protocol that uses public coins and shared objects, like common coins.

Metadata
Available format(s)
PDF
Category
Applications
Publication info
Preprint.
Keywords
blockchainbyzantine fault tolerant protocolsdistributed computingDAG
Contact author(s)
nassarshafik @ gmail com
History
2023-08-07: revised
2023-02-14: received
See all versions
Short URL
https://ia.cr/2023/192
License
Creative Commons Attribution
CC BY

BibTeX

@misc{cryptoeprint:2023/192,
      author = {Hagit Attiya and Constantin Enea and Shafik Nassar},
      title = {Faithful Simulation of Randomized BFT Protocols on Block DAGs},
      howpublished = {Cryptology ePrint Archive, Paper 2023/192},
      year = {2023},
      note = {\url{https://eprint.iacr.org/2023/192}},
      url = {https://eprint.iacr.org/2023/192}
}
Note: In order to protect the privacy of readers, eprint.iacr.org does not use cookies or embedded third party content.